Experian Decision Analytics - e-news February 2009

Client Focus

Debate

News & Features

Subscribe to e-news bi-monthly by email.
Click here »

Adapting credit strategies to survive and thrive

Given the current climate, many could be excused for feeling as if the sky were about to fall in. A question a few of the more pessimistic finance professionals are asking is whether credit scoring still works.

Will scorecards continue to separate the good risks from bad, and can they predict those consumers that may look good now, but are more likely to turn delinquent in a downturn? Experian’s analysis of its own bureau scores revealed that these scores continue to be strongly predictive of credit risk. We have also seen some evidence of an increase in delinquency levels at lower scores which is consistent with anecdotal evidence from a number of our clients. As the economic climate worsens people with it will be the people with the higher debt burdens and some history of previous credit problems who will be most vulnerable. These are the people to tend to score in the lower ranges.

The combination of rising unemployment and a decrease in real disposable incomes is likely to cause greater numbers to go delinquent on their credit commitments. However, existing credit scoring methodologies continue to be effective risk management tools in identifying those most likely to default. Increasingly, we see forward thinking financial institutions already starting to optimise their credit risk tools for further competitive advantage and see considerable value in the deployment of sophisticated decision analytics tools.

Credit scorecards rely on the assumption that the past predicts the future. They are built using data taken from a point, or perhaps a relatively short, period in the past. Historically, these scorecards have not included any direct measures of economic performance and can be thought of as being backward looking. The current economic climate has provided an incentive for some organisations to look at how to make scorecards more forward looking by factoring what is happening in the macro-economy into credit scoring models. If successful, this will enable scorecards to become more effective because they will include predictions about the future state of the economy. There are a number of technical challenges in doing this but economic necessity will no doubt be the mother of statistical invention. Despite some people’s concerns around the effectiveness of scoring, a return to the days of largely manual underwriting is not a sustainable answer.

''The application of risk management technology is more important now than ever before...''

Credit risk strategies can be optimised by looking afresh at scorecard monitoring and maximising the use of all of the available internal and external data. Organisations can still use automation to make more appropriate and accurate decisions. It is now more important than ever that financial institutions are proactive in monitoring the performance of their scoring models. Scoring models do degrade over time, and there are still relatively sophisticated financial institutions that do not monitor the performance their scoring models as effectively as they might. While there are no hard and fast rules about the life expectancy of scorecard, in more stable times one could expect it to be effective for two-three years.

If monitoring suggests that significant changes have occurred in the portfolio then the monitoring system should be able to produce the analysis necessary to quantify the scale of the change and its impact on the scorecards. This will inform the decision about whether scorecard changes are necessary. All scorecards should be monitored and the information reviewed at least every three months.

Maximising the use of internal and external data is an important part of risk management. Now is a good time to review the use of external data to ensure that all relevant data is being taken: organisations need to ask themselves whether they are making full use of the available credit bureau information and do they have access to the latest indebtedness and affordability measures if such things are available in the market. Too few organisations are incorporating measures of indebtedness into their scoring models. It doesn’t take a rocket scientist to figure out that the more indebted an individual is, the higher the risk is that they will become delinquent in the future.